top of page

Choosing a sovereign service provider—does it protect against cloud service outages?

4 hours ago
3 min read

Choosing a sovereign service provider—does it protect against cloud service outages?

Global outage at Microsoft 365, incident at AWS EC2, outages at Azure Front Door, Cloudflare, and OVHcloud: The incidents that recently disrupted the operations of cloud services for private and public companies raise a major question: Does choosing a sovereign service provider offer protection against service interruptions? The answer is no. The European DORA and NIS2 regulations, however, clearly outline the obligations of companies affected by outages in their business continuity plans.



These incidents reveal a common thread: business continuity. When a provider performs services that are essential to a large number of organizations, even the slightest failure in its infrastructure or cloud services can trigger a chain reaction of disruptions. Customers have no choice but to wait for their provider to restore service, risking business losses or productivity declines.


The cloud isn't infallible.


Following the global outage that affected Microsoft’s business email service, Office 365 applications, and the millions of users who rely on them, a question has been raised: Could choosing a sovereign provider avoid this chaos?


The answer is no. First, because the cause of this outage was related to a process responsible for establishing secure connections on Exchange servers. Once these servers became overloaded, they stopped responding. The load-balancing mechanisms then logically redirected connections to the servers that were still available, which themselves quickly became overloaded and unable to respond.


At AWS, the problem came from the DNS configuration, which impacted the EC2 virtual server infrastructure. At Cloudflare and Azure Front Door, the problem was a configuration issue. All of these issues led to the unavailability of services.

Cloud providers are not infallible. SLAs are essentials, and companies are responsible for ensuring their resilience in the event of outages or malfunctions.

When it comes to service availability, it is worth noting that most cloud providers offer global service infrastructures whose availability must be covered by SLAs (Service Level Agreements).These SLAs vary from one provider to another. They are based on a Shared Responsibility model (providers vs. users).

Microsoft could, for example, in the event of an Outlook and Exchange outage, become liable for the ineffectiveness of its load balancing. However, this does not exempt the user company from ensuring that its load balancing is properly configured. This is what makes this liability model so complex.


Sovereignty as a solution?


The key issue arising from all these incidents is therefore not the provider’s origin or the location of the servers, but rather the guarantee of service availability, their resilience, and the business continuity plan in place. The choice of a sovereign cloud provider does not affect the shared responsibility model or the need—or even the obligation, for certain organizations—to implement an appropriate business continuity plan.


Business Continuity and Redundancy: The Real Challenges.


It would be a mistake to forget that cloud security must be based on a governance approach—specifically, the implementation of resilient architectures that ensure business continuity in the event of an outage. Relying on a single “sovereign” cloud provider that could fall victim to an outage—such as the one that affected OVHcloud—would not fundamentally solve the problem.


The only way to avoid a situation where services essential to your organization are completely unavailable is to apply the principle of “not putting all your eggs in one basket” and reduce your technological dependence on a single provider.


This brings us back to a “best-of-breed” approach (selecting the best service for each critical or non-critical function), choosing two providers for essential services, and verifying the ability to easily switch from one to another to ensure operational continuity, using for instance open and standard APIs, open source stack, etc.


How does legislation regulate resilience in the event of a system failure?


1

DORA

Financial institutions must identify, classify, and continuously monitor their information systems, critical functions, and related services.

They must test the resilience of their information systems against cyber risks and system failures (vulnerability assessments and scans tests)

Contracts with suppliers must include rigorous clauses regarding SLAs, reversibility, auditability, and security..

Institutions affected by the Microsoft outage must prove to financial regulators (such as the ACPR in France) that they had alternative communication channels and operational continuity plans in place to maintain their critical operations during the outage.

2

NIS 2

The organizations concerned must implement backup, disaster recovery (DR), and business continuity plans.

They must assess and ensure the security of suppliers who have access to their critical systems.

For the financial sector, the DORA regulation prevails (Lex specialis)

An incident that causes a service interruption lasting more than one day across Europe triggers the requirement to report it to the relevant national security authorities (ANSSI) within 24 hours. Hospitals, public administrations, and private messaging companies must document their crisis management efforts as part of their business continuity obligations.






bottom of page